Project Summary Most research models of drug resistance are developed in the lab from cancer cell lines; however, their derived biological history is distinct from what occurs in patients. With the emergence of innovative methods to develop patient-derived models of cancer (PDMC), new opportunities are available to establish a diverse array of models representing individual patients? clinical experiences ? enabling identification of the features of tumors that underlie drug response and resistance in patient tissues. We hypothesize that PDMCs can be used to identify effective new therapies for breast cancer, and that analysis of cohorts of responding and non- responding PDMCs can be used to identify markers that predict treatment response or resistance. The goal of this proposal is to utilize a large and diverse bank of breast PDMCs to screen for new therapies for breast cancer, with a focus on current NCI-IND agents. We will first use patient-derived organoids (PDOs) to conduct high throughput screens of NCI-IND drugs as single agents and in combinations of two agents. Each drug and drug combination will be tested on PDOs from high-risk and metastatic breast tumors from 100 different patients. Effective ?lead? therapies will then be validated in patient-derived xenograft (PDX) models derived from a subset of those patients, to include predicted responders and non-responders based on PDO assays. Importantly, genomic data that is available on all of the PDX models will be utilized to identify proteomic, gene expression and/or mutational biomarkers associated with responders or non-responders, for each lead therapy. The primary outcome of this study will be to identify new NCI-IND treatment regimens for breast cancer using PDMCs as a pre-clinical tool to (1) evaluate the efficacy of many drugs and drug combinations across 100 different patients? tumor models, and (2) identify the features of responding and non-responding breast cancers for ?lead? therapies, in order to inform future clinical trial design.